Given we want Altair to be the the testing ground for Centrifuge and want everyone running Centrifuge infrastructure to also be familiar with Altair and help us spot issues there first, wouldn’t it be a good idea and easy requirement to:
- Require any active collator on Centrifuge to be active on Altair as well?
- Require any collator candidate for Centrifuge to be active for at least 1 month on Altair before being added to Centrifuge?
What do you think? I believe this would make both networks more secure and should be a reasonable ask for everyone involved.
And what to do with the collators that operate on Altair, but are not present in Centrifuge (Staking4all, Cardinate) and vice-versa?
Does this mean that number of total collators should be increased from 10 to 15?
There obviously would have to be a transition period (e.g. 3mo) in which all collators who aren’t on altair are asked to set up an altair collator. For collators who are on Altair today, they would be considered new collators and could be added to Centrifuge whenever they are ready but no later than a given time.
I don’t have an opinion on the ideal technical collator count, perhaps someone on engineering could weigh in on this.
One argument I have left out why I think this would be a good idea: it’s challenging for collators on Altair to make sufficient amount in AIR rewards to pay for the infrastructure necessary however I don’t think we should increase AIR rewards for collators. This proposal would effectively make sure that collators can use Altair as their test infrastructure for Centrifuge and gives them a way to test out new releases while at the same time ensuring they have the economic incentive to do so without relying on the proceeds of their AIR rewards.
My suggestion is to leave everything as it is, where individuals can choose to participate in either Altair or Centrifuge chains, providing flexibility and autonomy to the participants.
If someone joins Altair initially and demonstrates good performance over time, it’s fair to offer them the opportunity to become a collator in Centrifuge as well. The key condition for admission to the second chain should be the continued support of their collator’s work in Altair.
Similarly, those who are collators in Centrifuge should be encouraged to start a collator role in Altair within a specified timeframe.
Regarding rewards in Altair, I agree that monetary incentives are essential to motivate most people. Compensation should cover the expenses involved in the work, and ideally, leave some extra for personal treats, like a chocolate bar! Fair rewards are crucial for the sustainability and growth of any venture.
Is the thinking that it will be tough to find additional operators to run Altair collators b/c of costs / low returns? Are there any interested operators in the queue who would like to run an Altair collator?
Without knowing the answers to the questions above, I’m more aligned with @n1trog3n’s sentiments. Keep things as they are and allow operators to choose which network they would like to contribute their time and resources. Additionally, ensuring Altair collator operators are fairly compensated seems justified imo. In another thread, we discussed revisiting the amount of AIR collators would receive every few months. This seems logical to me, as it would/should be for Centrifuge collators as well (although CFG seems to be at a price and trajectory that doing so probably [hopefully] won’t ever be necessary).
There are three reasons why:
- Altair is relatively illiquid as a token and having 10 collators sell even a few thousand AIR a months to cover operational cost probably is not a good idea. It will just send the token into a downward spiral.
- There’s a benefit of Centrifuge collators also running Altair; we make sure Centrifuge collators test Altair which effectively is a pre-release of Centrifuge.
- Collators get the benefit of scale: it should be more efficient to run two collators than one one and as such Centrifuge and Altair will both benefit.
Fair rewards are crucial for the sustainability and growth of any venture.
I agree - but I don’t think that’s feasible for Altair without the support of Centrifuge at this point.
The only thing I would say about point 3 is it requires additional resources. One might have the computing power to run both collators on one server, but perhaps not the disk space. That is the case with us. We would need to acquire additional disk space to run both collators on one server which of course increases monthly costs.
Just getting caught up on this topic, Lucas, and wanted to offer my thoughts:
I like the idea, but based on previous comments perhaps this should be a “moving forward” idea. This way the current collators who are not on both chains don’t necessarily have to run both (and there are only two collators who are not) and incur additional costs, but it can be a requirement for anyone looking to join the Centrifuge set in the future. Then the process can be: 1) collator demonstrates an interest/willingness to support Centrifuge; 2) collator agrees to run on Altair for a set period (90 days?); 3) if performance metrics are sufficient, then that collator can be added to the Centrifuge set.
Just some additional thoughts based on the discourse that has transpired over the last few days…
I would say that the biggest question is the position of AIR token in the foreseeable future. To be honest, $125 of AIR * 10 (we have 10 collators now and definitely not everyone would be selling now) shouldn’t be such a huge selling pressure for any project - and if it is, we have to ask what are plans are out there for Altair and what role does the AIR token play in it. It’s been almost two years since Altair won an auction and I don’t think it’s any collator’s fault if the market can’t handle around ~$1k of selling pressure monthly so we can pay for basic infrastructure. This is not a rant, I am just trying to find out whether this is going to be a long-term situation or if there’s any utility that would help AIR market to recover on the horizon.
We have discussed the size of the fair reward for weeks and we came to $125 as a compromise between covering operational costs and not asking too much - of course, this can be discussed again if the team desires but I still find it reasonable.
Few points to the aforementioned proposal
- Collator count between networks usually varies between 10 and ~70 so we are definitely on the lower end of the spectrum
- Requirement for Centrifuge collator to run Altair collator as well would be quite a precedent in the ecosystem - afaik no one has done this before. Actually projects that chose collators with governance usually do it the other way around (different collators on different networks) for the sake of collator set diversity.
- I am not sure what the economic incentive would be - increased rewards in CFG for running Altair as well? How would the system look like reward-wise?
Sorry for jumping into the discussion this late, I was on vacation last week.
I want to make one thing clear, this wasn’t supposed to kick of a discussion on how much Altair collators should get but rather my view on Altair as a whole: in the current market climate we have to focus our effort on Centrifuge and ensure that Altair is able to serve as a canary and test network for Centrifuge. If we look at Altair as a test network for Centrifuge then expecting Centrifuge collators to also run Altair infrastructure is beneficial because we want them to test the software before it gets deployed to Centrifuge.
We have no network today that a large amount of outside collators run besides Altair so we have to assume that Centrifuge collators run no test network at all and are deploying the new Centrifuge releases without any testing whatsoever which is risky.
That is my main argument for proposing this.
Given my experience running infrastructure in general, and now for a couple of years focused entirely on Blockchain and Centrifuge I wanted to give my two cents:
The way I see it we are looking at this from the perspective of “incentives to cover my costs” probably because no slashing mechanism has been implemented yet in our parachain, that in an on itself is an incentive to run our nodes. Otherwise we would be looking at this as “the incentive is to cover my loses”. So when slashing is implemented the idea is not to cover the costs of the node with AIR, but to cover “your ass” over in the next runtime or node upgrade you will have to go through by getting slashed (making mistakes) on Altair first, where your amount of $(USD equivalent) lost will be a lot less than if you ran into the same issues on Centrifuge.
Now with the tokens at this price I can understand the general feelings here. We are all contributing to a still-growing network for a small incentive with the hope and possibly expectation that the token(s) will be more valuable in the future and you will be a privileged holder of them thanks to your staking/block producing efforts. That’s great! Thank you so much for your trust in our team and our project, we also believe in it with our full hearts!
But in this case, why not go the extra mile and run an Altair Node?
Contribute to the network somewhat altruistically today and when slashing is implemented you’ll be the most prepared node operator in the network, this should also bring you closer to our development team already by knowing about our upgrades, new functionalities, etc sooner. And in general, it is a good practice in our SW development industry to be able to try out your software upgrades and operations before hand in a close-to-production environment and that is what Altair is for.
But we all know the world doesn’t run entirely on good faith thus while this idea is great, until we figure out slashing I do think it could be interesting to ask all active collators to contribute to Altair. Maybe the incentive could be X months free of slashing your funds (or any other honesty mechanism) for “original/fist” collators in the network once its implemented.
On the other two topics on this thread:
- Collator count: The ideal collator count is no less than 5 (knowing K/F runs 3). But technically speaking a single honest collator should keep the chain happy and running (slow, yes). But if we get into the rabbithole the important idea is that at least 2-3 nodes are located far from each other, in completely separate providers and data centers, more than a certain number. 1 in Japan, 1 in US, and 1 in Europe in completely different providers is much much better than 10 nodes all running in AWS us-east1.
- Running on Altair first: I don’t think it is 100% necessary to “run for 3 months in Altair”, it really depends if the operator has 3 gold medals running PoW nodes already (say Onfinality for example) or if it is a totally new player nobody knows about? Just thinking out loud because I would honestly have to dig deeper in our current collator process before to understand this idea.